2024 — Scientific Reports

Detection of atmospheric radon concentration anomalies and their potential for earthquake forecasting using Random Forest analysis

Mayu Tsuchiya*, Hiroyuki Nagahama, Jun Muto, Mitsuhiro Hirano & Yumi Yasuoka
Random Forest machine learning was applied to atmospheric radon concentration data to detect anomalies potentially associated with earthquakes. The method achieved improved detection accuracy compared to conventional approaches, demonstrating that machine learning can enhance the identification of precursory radon signals for earthquake forecasting research.
2021 — Scientific Reports

Radon degassing triggered by tidal loading before an earthquake

Yasutaka Omori, Hiroyuki Nagahama, Yumi Yasuoka & Jun Muto
Anomalous atmospheric radon concentration changes before an earthquake were found to be associated with tidal loading. A physical model linking crustal strain from tidal forces to enhanced radon degassing from the ground was proposed, providing a mechanistic link between tidal triggering and precursory radon signals.
2021 — Scientific Reports

Preseismic atmospheric radon anomaly associated with 2018 Northern Osaka earthquake

Jun Muto, Yumi Yasuoka, Nao Miura, Daichi Iwata, Hiroyuki Nagahama, Mitsuhiro Hirano, Yoshiro Ohmomo & Takahiro Mukai
A significant decrease in atmospheric radon concentration was detected before the 2018 Northern Osaka earthquake (M6.1). Statistical analysis confirmed this anomaly as a precursory signal, supporting the hypothesis that crustal strain changes before earthquakes can modulate radon emanation from the ground surface.
2018 — Scientific Reports

Non-parametric detection of atmospheric radon concentration anomalies related to earthquakes

Daichi Iwata*, Hiroyuki Nagahama, Jun Muto & Yumi Yasuoka
A non-parametric anomaly detection method based on singular spectrum transformation was applied to atmospheric radon concentration time series. The method successfully identified anomalous radon changes that preceded several earthquakes, providing a statistical framework for earthquake precursor detection without assuming specific signal models.